Here are some indisputable facts about the challenge of assessing the ‘efficacy’ – in particular, benefits and risks, of Covid-19 vaccines:
Without an independent clinical assessment or post-mortem we can never be sure that the vaccine was the main or even a contributory factor if a person suffers an ‘adverse reaction’ or death after receiving the vaccine.
All measures of ‘effectiveness’ used so far rely on comparing numbers of ‘Covid-19 cases’ of vaccinated v unvaccinated people where a ‘case’ is defined by a positive PCR test. But PCR tests are unreliable and so we can never be sure if a person does or does not contract Covid-19. Moreover, there is no uniformity in testing strategies between unvaccinated and vaccinated people.
Even if we could accurately measure the true number of Covid-19 ‘cases’ over time, we can never be sure of the main cause of any change in numbers (vaccines, lockdowns, heard immunity, seasonality, etc).
The notion of whether or not there are ‘excess deaths’ (or excess illnesses) depends on multiple subjective criteria about previous years’ of data which may be incomplete. Moreover as with trends in case numbers, we can never be sure of the main cause of ‘excess deaths’.
It follows that any study that claims to demonstrate benefits and/or risks of Covid-19 vaccines which assumes to know some or all of the following in its data is potentially flawed:
a death/illness was or was not caused by the vaccine
the difference in Covid-19 incidence for vaccinated v unvaccinated by using PCR testing
decreasing or increasing ‘cases’ was or was not caused by the vaccine
there is a known number of ‘average’ deaths that should be ‘expected’
Until we have longer term data (especially on deaths for both vaccinated and unvaccinated) these problems can only be avoided by more independent clinical assessment, and a much more rigorous vaccine reaction yellow card reporting system. If Covid-19 'cases' really need to be considered to evaluate effectiveness then only those where there is a clinical diagnosis with illness and symptoms should be counted. If Covid-19 is as dangerous as is generally assumed, then, in the longer term, the only criterion we need to determine if vaccines work and are safe is if the proportion of deaths and serious illnesses among the vaccinated is lower than that among the unvaccinated.
For the longer term analysis we can completely ignore: a) whether or not a death/illness is due to Covid-19; b) the notion of what is/not a Covid ‘case’; c) any previous years’ data. We just need to know the following national (or large area) data for a few fixed time periods (initially we could even ignore the serious illness data):
If the (proportional) number of deaths/illnesses overall are less in a vaccinated age group than the corresponding unvaccinated age group, then the vaccine is saving more lives than it is killing. It’s as simple as that. No more and no less.
A full cost-benefit analysis should also add the cost of the vaccination programme (e.g. by using standard formulas that equate amounts of spending on healthcare with number of lives saved).